paper-with-me

홈 › Papers

WaterDrum: Watermarking for Data-centric Unlearning Metric

2025-05-08 · Xinyang Lu, Xinyuan Niu, Gregory Kang Ruey Lau, Bui Thi Cam Nhung, Rachael Hwee Ling Sim, Fanyu Wen, Chuan-Sheng Foo, See-Kiong Ng, Bryan Kian Hsiang Low

Large language model (LLM) unlearning is critical in real-world applications where it is necessary to efficiently remove the influence of private, copyrighted, or harmful data from some users. However, existing utility-centric unlearning metrics (based on model utility) may fail to accurately evaluate the extent of unlearning in realistic settings such as when (a) the forget and retain set have semantically similar content, (b) retraining the model from scratch on the retain set is impractical, and/or (c) the model owner can improve the unlearning metric without directly performing unlearning on the LLM. This paper presents the first data-centric unlearning metric for LLMs called WaterDrum that exploits robust text watermarking for overcoming these limitations. We also introduce new benchmark datasets for LLM unlearning that contain varying levels of similar data points and can be used to rigorously evaluate unlearning algorithms using WaterDrum. Our code is available at https://github.com/lululu008/WaterDrum and our new benchmark datasets are released at https://huggingface.co/datasets/Glow-AI/WaterDrum-Ax.

📄 PDF Abstract BibTeX arXiv:2505.05064

Code (1)

lululu008/waterdrum 공식 구현 pytorch

Tasks

Large Language Model

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Invisible Watermarks, Visible Gains: Steering Machine Unlearning with Bi-Level Watermarking Design

2025-08-13 · Yuhao Sun, Yihua Zhang, Gaowen Liu, Hongtao Xie 외 arxiv

With the increasing demand for the right to be forgotten, machine unlearning (MU) has emerged as a vital tool for enhancing trust and regulatory compliance by enabling the removal of sensitive data influences from machin…

Image ClassificationImage Generation

Unlearning in LLMs: Methods, Evaluation, and Open Challenges

2026-01-19 · Tyler Lizzo, Larry Heck arxiv

Large language models (LLMs) have achieved remarkable success across natural language processing tasks, yet their widespread deployment raises pressing concerns around privacy, copyright, security, and bias. Machine unle…

Community-Centric Graph Unlearning

2024-08-19 · Yi Li, Shichao Zhang, Guixian Zhang, Debo Cheng

Graph unlearning technology has become increasingly important since the advent of the `right to be forgotten' and the growing concerns about the privacy and security of artificial intelligence. Graph unlearning aims to q…

Benchmarking Federated Machine Unlearning methods for Tabular Data

2025-04-01 · Chenguang Xiao, Abhirup Ghosh, Han Wu, Shuo Wang 외

Machine unlearning, which enables a model to forget specific data upon request, is increasingly relevant in the era of privacy-centric machine learning, particularly within federated learning (FL) environments. This pape…

BenchmarkingComputational EfficiencyFederated LearningMachine Unlearning+1

Fast, Secure, and High-Capacity Image Watermarking with Autoencoded Text Vectors

2025-10-01 · Gautier Evennou, Vivien Chappelier, Ewa Kijak arxiv

Most image watermarking systems focus on robustness, capacity, and imperceptibility while treating the embedded payload as meaningless bits. This bit-centric view imposes a hard ceiling on capacity and prevents watermark…

Semantic Communication